Flow maps are powerful tools for visualizing the direction and volume of movements for many different variables, including commodities, services, commuters, armies, bird migration, maritime traffic, telecommunications traffic, and financial transactions. Although flow maps are useful and aesthetically pleasing communication tools, few computational techniques exist for creating flow maps that are both explicit and easy to read. This research project will identify design principles and develop algorithms to facilitate the creation of easy-to-read flow maps with curved flow lines. The design of a computational technique for creating flow maps and the distribution of the new technique as open-source software will enhance capabilities for users to create effective and informative flow maps, which will increase the geographic literacy of specific users as well as the general public. This project will contribute to the development of methods for visual analysis of various types of movement data in space and time that often are voluminous. The project will contribute to cartographic theory and practice by ascertaining and articulating the design principles underpinning the creation of effective flow visualizations, and it will provide interdisciplinary education and training opportunities for graduate and undergraduate students. The open-source software modules to be developed during the course of this project will make interactive and animated flow maps easier to produce and interpret than flow representations created with current digital tools.
This research project will concentrate on origin-destination flow maps with curved flow lines, where the specific geometry of the paths between pairs of origin and destination locations is either unknown or unimportant. Research in information visualization and graph drawing has resulted in techniques for creating flow diagrams, but the specific requirements for geospatial flow maps have received little attention. Despite the long history and aesthetic qualities of flow maps, there is little scientific knowledge concerning the effective design of flow maps. The investigators therefore will develop greater understanding of the effectiveness of traditional design principles for flow maps, and they will convert these principles into computational techniques for creating origin-destination flow maps. They will conduct content analyses and human-subject experiments for evaluating design techniques used in manually designed flow maps. They will identify the manual design principles that are found to be most relevant and useful and transform these into a multi-criteria cost function, and they will derive a map layout using heuristic optimization methods. The resulting techniques will be verified in a user study to assess efficiency and interpretation accurateness.
